Echolocation Call Characteristics of Eastern US Bats
uppr lwr slp total species ƒ hi ƒ lo ƒ ƒ dur special characteristics c maxE slp slp @Fc slp
Longer calls (>5ms) typically display a strong inflection point at the knee, 45.7 79.5 41.8 48.2 7.2 11.5 2.0 2.4 4.8 pronounced downward tail Myotis 47 91 44 52 8.5 15 3.1 4.2 7.6 ending call, and an extended grisescens call body with broad 44 68 40 44 5.8 8.3 0.9 0.5 2.0 gray bat amplituded distribution. 41-51 53-107 37-46 41-85 2.4-10 3.6-29 0.5-12 0.0-13 1.3-20 Shorter calls (3-5 ms) are typically at a higher frequency than other geographically overlapping Myotis.
FM sweep a smooth curve (i.e., no inflection), beginning steeply and then increasing in Myotis leibii curvature*. May have a well 44.3 95.1 40.6 49.1 3.2 33.5 9.6 8.9 16.9 defined downward tail. Peak eastern 46 104 42 52 3.9 40 12 12 22 power of call typically small– 5 42 86 39 46 2.5 27 7.0 5.5 12 persists for at least 1 ms on 0 footed non–saturated calls. Forage 38-48 55-115 31-44 40-71 1.7-5.3 6.9-48 2.5-22 0.0-28 4.6-36 myotis close to ground or vegetation. *some calls may have an inflection, but the smoothly curved variant is diagnostic. t o Myotis FM sweep a smooth curve austro- 43.6 84.3 39.6 46.4 4.6 17.6 6.1 6.6 9.7 (usually no inflection), riparius beginning steeply and then 45 95 41 48 5.5 22 8.6 11 15 increasing in curvature*. May south- 42 73 38 44 3.8 13 3.6 2.2 4.7 have a well defined downward tail. Peak power of call 4 eastern 38-48 66-116 31-44 42-65 2.0-6.2 5.9-31 1.8-14 0.0-22 4.0-26 myotis typically persists for at least 0 1 ms on non–saturated calls.
Myotis Calls may have up to 100 kHz M septen- 43.2 104 37.0 51.3 3.9 24.2 11.7 13.1 18.6 of bandwidth. FM sweep may y trionalis 47 114 42 62 4.6 30 16 18 24 be nearly linear making ƒc 40 95 32 41 3.1 18 7.4 8.0 14 difficult to recognize. Quiet but o northern consistent calls. Fly near t long-eared 32-53 60-12 25-50 37-95 1.7-6.6 8.5-55 3.0-36 0.0-37 6.5-43 vegetation, often with a linear i myotis flight path when searching. s Distinctive longer call type (>4.5 ms) may have a secondary inflection leading into a 40.8 80.9 37.5 44.0 5.8 16.8 4.6 2.6 7.1 “ledge” or flat section <1.3 Myotis ms just prior to terminal sweep sodalis 42 90 40 47 6.6 21 5.8 4.6 9.2 39 72 35 41 5.0 13 3.3 0.5 5.1 or “tail.” Note: some Mylu long Indiana bat calls share this feature. 34-47 50-115 25-43 37-70 1.9-7.8 4.1-42 1.0-16 0.0-14 2.3-23 Distinctive shorter call type also has ending ledge, but ~5–15% of shorter Myle & Mylu also exhibit this feature.
Sometimes with multiple power Myotis 39.7 69.4 36.5 43.4 5.8 10.5 3.5 4.1 5.0 centers making calls look lucifugus 41 78 38 47 6.7 14 4.6 6.2 6.7 clumpy. Longer duration calls little brown 38 61 35 40 4.9 6.7 2.3 2.0 3.4 recorded in open air are more bat discriminating. Dur >7 and 34-46 47-104 27-43 38-73 2.0-7.8 3.0-37 1.0-15 0.0-17 2.2-23 Lwr slp <3 distinctive.
Humboldt State University Bat Lab, March 2011 1 Echolocation Call Characteristics of Eastern US Bats
uppr lwr slp total species ƒ hi ƒ lo ƒ ƒ dur special characteristics c maxE slp slp @Fc slp
Strongly inflected, almost vertical FM changing to low Perimyotis 42.6 57.6 41.3 43.9 7.1 7.7 1.1 0.2 2.5 slope below 47 kHz for the subflavus majority of the call. Calls 44 67 43 46 8.4 14 1.7 0.7 4.0 generally consistent across a 41 48 40 42 5.8 1.7 0.4 -0.4 0.9 Tri-colored sequence. Appear hockey bat 36-47 41-106 34-46 36-50 3.5-12 0.3-38 0.0-4.9 0.0-4.2 0.1-12 stick–shaped in sonogram when FM sweep is present. Some calls exhibit “squiggles.” 4 0 40.4 67.6 40.2 43.8 6.8 10.0 2.0 0.6 4.4 U-shaped calls; up–turn at Lasiurus end of call; may exhibit borealis 44 81 43 49 9.1 16 3.2 1.6 7.1 variable ƒc across sequence. 37 54 37 39 4.6 4.4 0.7 -0.4 1.7 Power smoothly centered in n Eastern Red call. Low frequency can go as Bat 29-49 29-99 28-48 29-73 3.2-16 0.1-25 0.0-10 0.0-8.1 0.1-17 o low as 30 kHz. n (In progress) M U-shaped calls; up–turn at 40.4 62.8 39.9 42.8 7.6 7.9 1.5 0.4 3.3 end of call; may exhibit y Lasiurus variable ƒc across sequence. o seminolus 44 76 44 48 9.7 13 2.4 0.9 5.3 Power smoothly centered in 36 50 36 37 5.5 3.0 0.6 -0.2 1.3 call. Low frequency can go as t Seminole low as 30 kHz. 33-49 38-87 36-44 35-52 4.9-11 0.7-17 0.4-3.5 0.0-2.3 0.6-8.4 i bat Possibly acoustically S indistinguishable from L. borealis. Sweeping curved calls that may 37.8 63.0 36.1 40.0 6.6 12.5 2.3 1.2 4.9 Nycticeius lack any inflection. Calls have more slope in body (lower humeralis 40 78 38 43 9.4 20 3.7 2.5 7.9 slope) than do similar-shaped 36 48 34 37 3.8 4.7 0.9 0.0 1.8 shorter and longer Pisu calls. Evening bat 31-43 35-101 28-43 32-48 3.3-14 0.5-32 0.0-6.1 0.0-6.5 0.1-13 Sequences may display ƒc alternating up and down.
How to use this table This table presents ranges for the general characterizing call parameters of echolocation calls. The boldface numbers display the mean and ± standard deviation of the subset of calls correctly identified using SonoBat automated classification, i.e., the subset of each species call repertoire having the most species-discriminating characteristics, using the default 0.90 discriminant probability threshold. The lower, smaller font numbers display the overall range of all calls in the library of species-known calls used to prepare the SonoBat classifiers. Bold text indicates the most species–discriminating characteristics. Because of intraspecific variablity and similarity with other species, the parameters presented here will often be insufficient for confident identification. SonoBat extracts more than five dozen parameters that it uses for call and sequence classification. Analyze 1) well–formed calls, i.e., search phase calls recorded from bats in a steady mode of flight, away from roosts and not accelerating or performing some other maneuver that elicits rapid, short calls, e.g., like that from a hand–released bat, and 2) calls with a strong signal that clearly rise above the background noise level and have little distortion or echoes. It is generally preferable to avoid analyzing calls that saturate, i.e., overload, the detector or recorder. However, saturated call specimens may still be used to interpret time-frequency characteristics, but consider the time-amplitude domain from saturated calls to be unreliable. Terminology and key lo ƒ: lowest apparent frequency (kHz), hi ƒ: highest apparent frequency (kHz); this can vary depending upon the distance to the bat, ƒc: characteristic frequency, i.e., the frequency of the call at its lowest slope toward the end of the call, or the lowest frequency for consistent FM sweeps (kHz), ƒmaxE: the frequency with the greatest power (kHz), dur: call duration from the beginning to the end of the call (ms), upper: the slope of the upper portion or onset of the call (kHz/ms) from the high ƒ to the knee (listed as HiFtoKnSlope on SonoBat output), lower: the slope of the lower portion or body of the call (kHz/ms) from the knee to the ƒc (listed as KnToFcSlope on SonoBat output). FM: frequency modulation, i.e., a change in frequency with time, flat: a call or portion of a call with a very low slope or no slope (horizontal), inflection: a pronounced change in the slope of a call, sometimes called a “knee,” power: the amplitude or sound energy of a call or portions of a call, squiggle: an S-shaped variation in frequency with time over a portion of the call.
Humboldt State University Bat Lab, March 2011 2 Echolocation Call Characteristics of Eastern US Bats
uppr lwr slp total species ƒ hi ƒ lo ƒ ƒ dur special characteristics c maxE slp slp @Fc slp
U-shaped calls; up–turn at end of call; may exhibit variable ƒc across sequence. Lasiurus 28.4 45.5 27.9 29.4 10.5 4.9 0.9 0.06 1.9 Power smoothly centered in intermedius 30 53 29 32 13 7.3 1.4 0.3 2.9 call. Low frequency can go as northern 27 38 26 27 8.3 2.4 0.4 0.0 0.9 low as 25 kHz. Calls similar in yellow bat shape and variability to other 33-25 29-79 25-32 25-41 3.7-16 0.5-14 0.2-2.9 0.0-2.4 0.3-6.0 Lasiurans, but intermediate in frequency range between Labo/Lase and Laci. Variable; calls with high ƒ below 60 kHz can be confused with Lano and/or Tabr. Calls with high ƒ above 65 kHz 27.9 49.4 26.5 30.0 8.2 5.8 1.8 1.5 3.1 distinguish from Lano, even Eptesicus long calls have some FM fuscus 30 56 28 32 11 8.3 2.9 2.9 4.6 component, i.e. never flat. The end of calls may hook upward. big brown 26 42 25 28 5.3 3.3 0.7 0.1 1.5 bat 21-33 29-69 19-32 22-42 2.8-19 0.9-17 0.2-9.4 0.0-8.2 0.3-12 * Shorter calls recorded with 3 full detail, i.e., ones that closely approached the 0 microphone, as indicated by the presence of harmonics, exceed 65-70 kHz. Shorter calls reverse J–shaped; often with a distinct inflection. Some call variants can be confused with Epfu and/or Tabr. Flat calls ≥26 kHz diagnostic. Flat Laci calls are lower in ƒ. Lasio- 26.6 41.7 25.4 28.8 8.9 5.5 1.3 1.1 2.6 Low slope calls in the 25–26 nycteris kHz range may be distinguished noctivagans 28 51 27 31 13 9.3 2.7 2.6 4.9 from Laci by the presence of an 25 33 24 27 4.8 1.7 0.0 -0.5 0.4 inflection. Epfu has more FM, silver– typically with smooth curvature 23-31 26-63 14-30 24-44 2.3-24 0.0-22 0.0-8.8 0.0-8.3 0.0-12 haired bat (no inflection). * Shorter calls recorded with full detail, i.e., ones that closely approached the microphone, as indicated by the presence of harmonics, still do not exceed 50-55 kHz.
Caveats: Please note that the range of characteristics listed in this table overlap among many of the species, and that although compiled from over 11,000 calls, it still represents a finite, noninclusive data set and any individual bat may emit calls beyond the typical ranges and call characteristics listed in this table (and mimic another species). This and the variability of bat echolocation calls renders acoustic classification of bats a probabilistic process and relatively inexact compared to a process like genotyping. For some species, confident species classification can only be achieved on a subset of call types within its repertoire that falls outside of data space shared with another species. As a result many recordings will have ambiguous species classification. Expect that, and seek the most species-discriminating call types on which to make species determinations.
Species classification also depends upon accurate extraction of call parameter data, and that depends upon high quality recordings having clear signals with a high signal to noise ratio and free from distortion and confounding echoes.
Humboldt State University Bat Lab, March 2011 3 Echolocation Call Characteristics of Eastern US Bats
uppr lwr slp total species ƒ hi ƒ lo ƒ ƒ dur special characteristics c maxE slp slp @Fc slp
Low intensity, difficult to detect; harmonics often present. Call– 23.4 42.5 21.4 31.1 4.6 7.1 4.9 4.2 5.0 shape simple linear FM Corynorhinus sweep, (sometimes with townsendii 26 45 23 34 6.3 13 6.6 6.5 6.5 upsweep at onset). fmax may Townsend’s 21 40 19 28 3.0 1.2 3.2 1.9 3.5 alternate between fundamental and second harmonic. big–eared bat 18-32 36-51 17-30 22-41 1.7-11 0.2-70 1.1-13 0.0-13 1.0-11 This species sometimes applies more amplitude in the second harmonic than in the first. Low intensity, difficult to detect; harmonics often present. Call– shape simple linear FM 22.8 39.8 22.5 33.2 2.6 6.2 7.4 6.7 6.7 sweep, (sometimes with Cornorhinus upsweep or flat tone at onset rafinesquii 25 42 25 37 5.1 8.6 8.9 8.9 8.1 before sweeping downward). Rafinesque’s 20 37 20 30 0.1 3.7 5.9 4.4 5.2 fmax may alternate between big–eared bat fundamental and second 23-23 40-40 22-22 33-33 2.6-2.6 6.2-6.2 7.4-7.4 6.7-6.7 6.7-6.7 harmonic. This species sometimes applies more amplitude in the second harmonic than in the first. 2 Variable; FM to flat; can be 0 25.5 32.3 24.1 28.0 11.5 1.6 0.5 0.4 0.7 confused with Epfu, Lano, or Tadarida Laci. Long calls that "turn on" brasiliensis 28 39 26 31 14 3.2 0.8 1.0 1.4 power rapidly with high free–tailed 23 25 22 25 9.5 0.0 0.1 -0.3 0.0 energy at beginning bat 18-33 19-61 17-33 18-46 3.5-20 0.0-17 0.0-4.5 0.0-4.1 0.0-4.8 (oscillogram carrot–like). Calls often upswing into call and downswing out of call. Pronounced or subtle U– shape; very flat calls may have slight downturn into call or upturn at end. Fully formed (i.e., good quality) calls never get 20.1 26.0 19.7 20.8 11.0 2.2 0.4 0.0 0.7 completely flat like a Tabr or Lasiurus Lano, but out of range cinereus 22 31 22 23 15 4.1 0.8 0.2 1.4 fragments can appear flat and 18 21 18 18 7 0.3 0.1 -0.1 0.1 mimic Tabr. The most flat calls hoary bat tend to be lower in ƒ than flat 16-32 17-58 16-31 17-49 4-26 0.1-14 0.0-5.7 0.0-4.6 0.0-8.3 Lano calls. Low ƒ may vary across sequence, power builds toward center then gradually declines. Short calls can be confused with Lano, Epfu, or Tabr.
Sources: Characteristics gleaned from recordings acquired by J.M. Szewczak, Humboldt State University Bat Lab (J.M. Szewczak, Aaron Corcoran, and Jean-Paul Kennedy), Patricial C. Ormsbee, USFS Pacific Northwest Research Station, and various contibutors to the Pacific Northwest Bat Grid, T.J. Weller, USFS Redwood Sciences Lab, and numerous other generous contributors.
The information presented in the table represents work in progress and is presented with the acknowledgement that it is unlikely to be the definitive description of these species’ acoustic characteristics. Please use accordingly.
Humboldt State University Bat Lab, March 2011 4